[wtallis]

For your first point, the difference will only ever be very small when the topic of discussion is something that is inherently associated with a particular person, such as a patch with a specific submitter. The phrasing of the former example is more passive-aggressive, but usually no less targeted.

For your second point, you said:

"If you don't appreciate their code, get rid of them. [...] tell them it's not working out and stop working with them."

That implies more than just rejecting bad patches, it implies rejecting the developer himself. That's extremely rare. The Linux kernel developers are very forgiving of mistakes: your patches will get rejected if they're bad, but they'll still get looked at until you establish a really bad track record of not learning from your mistakes.